Indeterminate domain 10 regulates ammonium-mediated gene expression in rice roots


  • Yuan Hu Xuan,

    1. Division of Applied Life Science (BK21 Program), Plant Molecular Biology and Biotechnology Research Center (PMBBRC), Gyeongsang National University, Jinju, South Korea
    Search for more papers by this author
  • Ryza A. Priatama,

    1. Division of Applied Life Science (BK21 Program), Plant Molecular Biology and Biotechnology Research Center (PMBBRC), Gyeongsang National University, Jinju, South Korea
    Search for more papers by this author
  • Jin Huang,

    1. Division of Applied Life Science (BK21 Program), Plant Molecular Biology and Biotechnology Research Center (PMBBRC), Gyeongsang National University, Jinju, South Korea
    Search for more papers by this author
  • Byoung Il Je,

    1. Division of Applied Life Science (BK21 Program), Plant Molecular Biology and Biotechnology Research Center (PMBBRC), Gyeongsang National University, Jinju, South Korea
    Search for more papers by this author
  • Jing Miao Liu,

    1. Division of Applied Life Science (BK21 Program), Plant Molecular Biology and Biotechnology Research Center (PMBBRC), Gyeongsang National University, Jinju, South Korea
    Search for more papers by this author
  • Soon Ju Park,

    1. Division of Applied Life Science (BK21 Program), Plant Molecular Biology and Biotechnology Research Center (PMBBRC), Gyeongsang National University, Jinju, South Korea
    Search for more papers by this author
  • Hai Long Piao,

    1. Division of Applied Life Science (BK21 Program), Plant Molecular Biology and Biotechnology Research Center (PMBBRC), Gyeongsang National University, Jinju, South Korea
    Search for more papers by this author
  • Dae Young Son,

    1. Department of Applied Biology, Gyeongsang National University, Jinju, South Korea
    Search for more papers by this author
  • Jeung Joo Lee,

    1. Department of Applied Biology, Gyeongsang National University, Jinju, South Korea
    Search for more papers by this author
  • Sung Han Park,

    1. National Academy of Agricultural Science, Rural Development Administration, Suwon, South Korea
    Search for more papers by this author
  • Ki Hong Jung,

    1. Crop Biotech Institute, Kyung Hee University, Yongin, South Korea
    Search for more papers by this author
  • Tae Ho Kim,

    1. Genomics Division, Department of Agricultural Biotechnology, National Academy of Agricultural Science (NAAS), RDA, Suwon, South Korea
    Search for more papers by this author
  • Chang-deok Han

    Corresponding author
    • Division of Applied Life Science (BK21 Program), Plant Molecular Biology and Biotechnology Research Center (PMBBRC), Gyeongsang National University, Jinju, South Korea
    Search for more papers by this author

Author for correspondence:

Chang-deok Han

Tel: +82 55 772 1356



  • Indeterminate domain (IDD) genes are a family of plant transcriptional regulators that function in the control of development and metabolism during growth. Here, the function of Oryza sativa indeterminate domain 10 (OsIDD10) has been explored in rice plants. Compared with wild-type roots, idd10 mutant roots are hypersensitive to exogenous ammonium. This work aims to define the action of IDD10 on gene expression involved in ammonium uptake and nitrogen (N) metabolism.
  • The ammonium induction of key ammonium uptake and assimilation genes was examined in the roots of idd10 mutants and IDD10 overexpressors. Molecular studies and transcriptome analysis were performed to identify target genes and IDD10 binding cis-elements.
  • IDD10 activates the transcription of AMT1;2 and GDH2 by binding to a cis-element motif present in the promoter region of AMT1;2 and in the fifth intron of GDH2. IDD10 contributes significantly to the induction of several genes involved in N-linked metabolic and cellular responses, including genes encoding glutamine synthetase 2, nitrite reductases and trehalose-6-phosphate synthase. Furthermore, the possibility that IDD10 might influence the N-mediated feedback regulation of target genes was examined.
  • This study demonstrates that IDD10 is involved in regulatory circuits that determine N-mediated gene expression in plant roots.


The indeterminate domain (IDD) family is composed of a group of genes encoding putative nuclear proteins that contain four zinc finger (C2H2, C2H2, C2HC, C2HC) motifs called the indeterminate (ID) domain (Colasanti et al., 2006). The ID domain originated from a maize ID1 that regulates the floral transition from vegetative growth (Colasanti et al., 1998). Recently, the functions of other IDD genes have been elucidated by forward or reverse genetic approaches. In Arabidopsis thaliana, MAG (magpie/IDD3) and JKD (jackdaw/IDD10) are involved in root fate determination, whereas SGR5 (shoot gravity response 5/IDD15) is involved in gravitropism. Recently, the A. thaliana genes IDD1/ENY (enhydrous), IDD8 and IDD14 have been shown to control metabolic processes for seed maturation and plant development (Morita et al., 2006; Welch et al., 2007; Feurtado et al., 2011; Seo et al., 2011a,b). Therefore, IDD genes are involved in the regulation of developmental phase progression and metabolic processes for plant growth.

In higher plants, nitrate and ammonium represent the major sources of nitrogen (N) for roots. Plants vary in their ability to utilize these two N forms. Although ammonium is an energetically favorable N source, many plants exhibit symptoms of toxicity when it is the sole N source (Britto & Kronzucker, 2002). The ammonium ion is the major N species in rice paddy soils and is utilized as the major source for N assimilation in rice crops. Ammonium is taken up directly from the soil by ammonium transporters (AMTs) and is assimilated into the amino acid glutamate via the glutamine synthetase/glutamate synthase (GS/GOGAT) cycle. N assimilation is linked to carbon and respiratory metabolism by the demands of the GS/GOGAT cycle for reductants and 2-oxoglutarate (2-OG) as a carbon skeleton, which necessitate the induction of enzymes in glycolysis and the Krebs cycle (Galvez et al., 1999). The major forms of organic N are glutamine and asparagine, which are transported from roots to shoots via the xylem (Fukumorita & Chino, 1982).

Systemic approaches have been used to explore genomic expression under N starvation or nutrient culture conditions in leaves and whole plants (Scheible et al., 2004; Lian et al., 2006; Zhu et al., 2006). Different N sources exert distinct effects on a substantial number of genes and physiological processes. In A. thaliana, nitrate exerts major effects on N metabolism and photosynthesis, whereas ammonium affects carbon and sugar metabolism, as well as respiration (Hoffmann et al., 2007). A series of transcriptome analyses have demonstrated that N assimilation is linked to various metabolic and developmental processes at the genomic expression level (Wang et al., 2000, 2003; Zhu et al., 2006; Gutierrez et al., 2008; Patterson et al., 2010). Although inorganic N or its metabolites can generate signals that induce transcriptional activity in metabolic and developmental genetic networks, only a few regulatory factors have been identified that are responsible for N-mediated gene induction and that link N signals to cellular and metabolic responses. The maritime pine transcription factor Dof has been shown to exert direct control on two GS genes via promoter binding (Rueda-Lopez et al., 2008), and the NIN-like transcription factor NLP7, which carries a DNA-binding domain, functions in nitrate-mediated induction of nitrate assimilation genes (Castaings et al., 2009). GATA transcription factors have been implicated in the regulation of N metabolism (Jarai et al., 1992), as has a regulator coordinating N assimilation and circadian rhythm, the CIRCADIAN CLOCK ASSOCIATED 1 (CCA1) gene, which encodes a MYB-related transcription factor (Gutierrez et al., 2008). The MADS-box transcription factor, ANR1, positively regulates the lateral root response to local nitrate supply, whereas LBD37/38/39 are negative regulators of nitrate signaling. In addition, two CIPK members, CIPK8 and CIPK23, are primary nitrate response genes in A. thaliana (Krouk et al., 2010; Castaings et al., 2011; Tsay et al., 2011).

In this study, the functional role of rice IDD10 was explored using mutants and overexpressors. The idd10 mutants exhibited ammonium hypersensitivity in root growth, which led us to examine the expression patterns of key ammonium uptake and N assimilation genes. The data showed that IDD10 activates the transcription of target genes via direct binding to cis-elements. Expression studies showed that many genes involved in N-linked metabolism are regulated by IDD10. Our findings suggest that IDD10 functions in regulatory circuits that control the expression of N-responsive genes.

Materials and Methods

Mutant isolation and plant growth

PCR screening identified a T-DNA insertion line (idd10) from a rice T-DNA population ( Primers for IDD10 (Os04g47860) and T-DNA were used (An et al., 2003). The mutant line was derived from Oryza sativa Japonica rice cv Hwayoung. Transgenic lines were generated from O. sativa Japonica rice cv Dongjin. The following growth conditions were used to examine the effects of ammonium on gene expression. After germination, plants were grown in distilled water in a glasshouse for 14 d. The seedlings were grown for a further 3 d in the N-free nutrient solution described by Abiko et al. (2005). Seedlings were then transferred to a nutrient solution containing 0.5 mM (NH4)2SO4 at pH 5.5. Whole roots were harvested at 0, 3, 6 and 12 h following the provision of (NH4)2SO4. To examine the effects of methionine sulfoximine (MSO) and glutamine, samples were grown for 3 d in N-free nutrient solution as described above, and then transferred for 3 h to the same medium containing 0.5 mM (NH4)2SO4, a mixture of 1 mM MSO and 0.5 mM (NH4)2SO4, or 1, 5 or 10 mM l-glutamine.

Transactivation assays

Transactivation assays were performed in the yeast strain PJ69-4A, which contains the lacZ and HIS3 reporter genes. Using the pGBT9 vector (Clontech,, DNA encoding the GAL4 DNA-binding domain was fused to the following IDD10 DNA fragments: the complete open reading frame (ORF), a 5′ cDNA encoding the first 255 amino acids or a 3′ cDNA encoding amino acids 255–402. The 3′ AtNAC1 fragment that encoded a peptide from amino acid 143 to 324 was used as a positive control. These constructs or the empty vector (pGBT9) were introduced into yeast cells. Yeast transformants were grown on yeast minimal media/synthetic defined (SD) premixes without tryptophan (SD/Trp−) and without histidine (SD/His−) (Rose et al., 1990).

Microarray analysis

Two microarray experiments were performed for the rice gene expression analysis. A preliminary microarray was performed with one biological replicate using four samples (idd10 mutant, IDD10 OX2 and their wild-type controls). In the second microarray, three biological replicates were conducted using four samples from a different OX line (idd10 mutant, IDD10 OX-GFP2 and their wild-type controls). Samples were grown as described above. After the provision of 1 mM ammonium for 0 or 3 h, 30 roots from each sample were harvested and pooled for the extraction of total RNA with the RNeasy Plant Mini Kit (Qiagen, All microarray experiments, including data processing, were performed according to the Agilent manuals at POSTECH Systems Biology Laboratory (Pohang, South Korea). For each dataset, we calculated fold-change values compared with 3 h treatment samples just before ammonium addition. The ratios of log2-normalized intensity values from the idd10 mutant or IDD10 OX were calculated against those of the corresponding wild-type samples. The data from the second microarray experiment with three biological replicates are presented in Supporting Information Table S1.

Real-time polymerase chain reaction (RT-PCR) analysis

Total cellular RNA was isolated with RNeasy Plant Mini Kits (Qiagen). The cDNA was synthesized using reverse transcriptase RNaseH (Toyobo, Quantitative RT-PCR products were quantified using the MJ and Illumina Research Quantity software (GeneEx Macro OM 3.0 (MJ-Biorad, Hercules, California, USA) and Illumina Eco 3.0, (Illumina, San Diego, California, USA) respectively), and values were normalized against Actin and Ubiquitin cDNA from the same samples. The primers used for RT-PCR are shown in Table S2.

Vector construction and cytological observation

For the construction of a green fluorescent protein (GFP) fusion vector, cDNA encoding the 1.2-kb ORF of IDD10 was isolated by PCR and fused to GFP. The PCR primers for cloning IDD10 cDNA and GFP are listed in Table S3. To create a complementation vector, a 2.5-kb fragment containing the IDD10 promoter was fused to the 1.2-kb full-length cDNA. For a β-glucuronidase (GUS) fusion vector, the 2.5-kb IDD10 promoter region was fused to the GUS reporter gene. Rice transformation, GUS and GFP detection, and in situ RNA hybridization were performed following standard methods (Jackson, 1992; Chin et al., 1999; Park et al., 2008).

Determination of ammonium contents

Enzymatic determination of ammonium content in the roots was performed using an F-kit (Roche), according to the manufacturer's instructions (Oliveira et al., 2002).

Chromatin immunoprecipitation (ChIP) assay

Eight grams of calli were prepared from Ubiquitin:IDD10:GFP and 35S:GFP transgenic plants for the ChIP assay, which was performed according to the published method (Ito et al., 1997; Je et al., 2010). Immunoprecipitates were analyzed by semiquantitative PCR. Each input DNA level was used as an internal control. PCR primers for the ChIP assay are listed in Table S3.

Electromobility shift assay (EMSA)

To produce glutathione S-transferase (GST):IDD10 fusion protein, IDD10 cDNA (67–254 amino acids) was subcloned into the pGEX 5X-1 expression vector (Amersham). For probe labeling, DNA fragments of 15–40 nucleotides were synthesized and then either end-labeled with [γ-32P]ATP using T4 polynucleotide kinase (NEB, Ipswich, Massachusetts, USA) or labeled with [α-32P]dATP by PCR. EMSAs were performed following a published method (Je et al., 2010).

Transient expression assay

An effector (pUbi:IDD10), reporters (pAMT1;2:GUS (Pwt) and mutated construct-GUS (Pm1) fusions) and an internal control (p35S:LUC) were cotransformed into rice Oc protoplasts (Wong et al., 2004). To generate pUbi:IDD10, IDD10 cDNA was cloned into the pUC19 vector containing a maize Ubi promoter (pUbi:PAT2). To generate the reporter vectors, the 2-kb AMT1;2 promoter was isolated by PCR. To generate Pm1, a mutated sequence (AAAAAAA) was introduced 244 bp upstream of the ATG codon of the 2-kb promoter. Wild-type and mutated promoters of AMT1;2 were cloned into the pBI221 vector containing the GUS coding region. To generate p35S:LUC, the 1.7-kb coding region of luciferase was fused to the 35S promoter in the pUC19 vector. GUS expression was normalized against luciferase expression. Electroporation and activity assays were performed as described previously (Wong et al., 2004). The PCR primers for cloning the normal and mutated AMT1;2 promoter are listed in Table S3.


idd10 mutants develop an ammonium-hypersensitive root phenotype

To elucidate the function of an IDD10 gene in rice, a null mutant was isolated from a T-DNA population using PCR and target-specific primers. The T-DNA was inserted into the second intron of the IDD10 locus. Although idd10 mutants did not show any visible phenotype in the field, strong growth inhibition and coiling of root tips were observed in the seminal roots of seedlings grown on basal Murashige–Skoog (MS) salts medium (Fig. 1a). The root phenotype did not develop when mutant seedlings were grown in distilled water or under dark conditions (Fig. S1a,b). To verify the idd10 root phenotype, > 10 independent complementation lines were generated by transforming the idd10 mutants with a T-DNA expression vector containing the 2.5-kb endogenous IDD10 promoter fused to the full-length cDNA (Fig. S1c). Because complementation lines showed normal growth of seminal roots on MS salts medium (Fig. S1d,e), it was concluded that the root phenotype resulted from the genetic lesion in IDD10.

Figure 1.

Ammonium-dependent inhibition of seminal root elongation of indeterminate domain 10 (idd10). Rice wild-type and idd10 plants were grown for 3 d in (a) 0.5 × Murashige–Skoog (MS) medium, or (b) modified 0.5 × MS medium, in which 10 mM KNO3 replaced 10 mM NH4NO3. The medium contained all MS components, except for inline image. (c) Plants were grown for 3 d in 0.1, 0.25, 0.5, 1.0 or 2.0 mM ammonium. (d) Graphic comparison of seminal root length in mutant (red symbols) and wild-type (blue symbols) plants. Error bars represent ± SE of at least six samples. Significant differences between wild-type and idd10 are shown (*, < 0.05; **, < 0.001; t-test). (e) Plants were grown in 0.5 × MS medium without (left panel) or with (right panel) 1 μM methionine sulfoximine (MSO) for 3 d.

MS salts comprise 13 chemical components. To identify which component of MS salts might be responsible for the mutant phenotype, seedlings were grown in modified MS salts medium, with each medium lacking one component. The growth defect of seminal roots was observed in mutant seedlings grown on all media except for that lacking 10 mM NH4NO3 (Fig. S2a). To separate the effects of ammonium and nitrate, mutants were grown in modified MS salts medium in which 10 mM NH4NO3 was replaced with 10 mM KNO3, 5 mM K2SO4 or 5 mM (NH4)2SO4. Only medium without inline image produced normal growth of seminal roots (Fig. 1b). The mutant phenotype of idd10 was observed on medium containing inline image, but not on medium containing inline image (Fig. S2b,c). These data show that the root phenotype of idd10 is dependent on ammonium.

It should be emphasized that the seminal roots of normal rice plants also exhibit the same growth defect when grown in high concentrations of ammonium (Hirano et al., 2008). In addition, the root phenotype induced by ammonium can be suppressed by the addition of MSO, which prevents ammonium assimilation by GS. Therefore, the possibility was examined that the phenotypic difference in seminal roots between idd10 and wild-type plants might be caused by a difference in ammonium sensitivity. To compare ammonium sensitivity, mutant and wild-type plants were grown in various concentrations of ammonium (Fig. 1c,d). Both developed growth retardation and tip coiling of seminal roots. However, mutants developed root defects at much lower ammonium concentrations than did wild-type plants. Seminal roots of idd10 mutants were examined further for responses to MSO. Growth defects were no longer visible in the seminal roots when mutants were grown in MS salts medium containing MSO (Fig. 1e). These results suggest that the root phenotype of idd10 mutants might result from enhanced sensitivity to ammonium.

IDD10 expression patterns in plants and transcriptional activity in yeast

IDD10 is a member of the IDD family, all of which contain a zinc finger domain comprising two C2H2 and two C2HC motifs (Colasanti et al., 2006). Figure 2(a) shows the sequence alignment between IDD10 and OsID1, a maize ID1 ortholog (Park et al., 2008). Because some IDDs have been identified as transcription factors, the transcriptional activity of IDD10 was examined using a yeast system. In the transcriptional activation assay, vectors expressed the yeast GAL4 DNA-binding domain fused to the full-length IDD10 coding region, the N-terminal ID domain or the C-terminal region. The activation domain of an A. thaliana NAC gene SND1 was used as a positive control (Xie et al., 2000). The N-terminal peptide, which contained the ID domain, showed no transcriptional activity, whereas the C-terminal domain exhibited strong transcriptional activation (Fig. 2b). In yeast, the complete IDD10 protein was less active than the C-terminal peptide alone. For the subcellular localization of IDD10, a vector encoding GFP fused to the C-terminal end of IDD10 was expressed in transgenic plants under the control of the Ubiquitin promoter. IDD10:GFP OX lines complemented the idd10 mutation (Fig. S3a). A strong GFP signal was detected exclusively in nuclei, which indicates that IDD10 is a nuclear protein (Fig. 2c).

Figure 2.

Nuclear localization and expression patterns of indeterminate domain 10 (IDD10). (a) Sequence alignment of OsIDD10 and OsID1. Identical and similar amino acids are shown in black and gray boxes, respectively. Blue and red horizontal bars indicate the putative nuclear localization sequence (NLS) and indeterminate (ID) domains, respectively. Asterisks in the red bar indicate core amino acids of the two C2H2 and two C2HC zinc fingers. (b) Transactivation activity of different regions of IDD10. DNA encoding the full-length (402 amino acids), N-terminal (255 amino acids) and C-terminal (147 amino acids) regions of IDD10 were fused to the GAL4 DNA-binding domain and transformed into yeast cells. AtNAC1 was used as a positive control. Histograms show the levels of β-galactosidase activity. (c) The left panel shows confocal microscopy of green fluorescent protein (GFP) overlapped with propidium iodide (PI) staining in a seminal root of IDD10-GFP transgenic plants. The right panel shows light microscopy of the same tissue. Bars, 20 μm. (d) Quantitative real-time polymerase chain reaction (qRT-PCR) was performed on mRNA from roots and leaves of 1-wk-old rice plants, shoot apices of 3-wk-old plants and flowers of 3-month-old plants. mRNA levels were normalized to that of Ubiquitin mRNA. Error bars represent ± SE of the means from three replicates. (e) Seventeen-day-old seedlings grown as described in the 'Materials and Methods' section were then transferred to medium containing no nitrogen (−N), 0.5 mM (NH4)2SO4, a mixture of 1 mM methionine sulfoximine (MSO) and 0.5 mM (NH4)2SO4 or 5 mM l-glutamine (Gln). After 3 h, qRT-PCR was used to measure IDD10 mRNA levels. Ubiquitin was used to normalize the mRNA level of each gene. Data represent mean values ± SE of three replicates. Significant differences from the untreated samples are indicated by asterisks (*, < 0.1; **, < 0.05; t-test). (f) IDD10 expression in roots was analyzed using transgenic plants in which β-glucuronidase (GUS) was expressed from an endogenous IDD10 promoter. Bars, 2 mm. (g) In situ hybridization was performed with cross-sections of roots. Mutants hybridized with an antisense probe were used as a control (right panel). Arrows indicate root primordia. Bars, 0.5 mm.

Quantitative real-time-PCR was used to examine IDD10 expression in roots, leaves, shoot apices and flowers. IDD10 mRNA was detected in all tissues, and particularly high expression levels were observed in flowers (Fig. 2d). The expression of IDD10 was suppressed by ammonium and glutamine. However, nitrate treatment had no statistically significant effect. MSO prevented the ammonium-mediated suppression (Fig. 2e). To investigate IDD10 expression patterns in roots, transgenic GUS and RNA in situ hybridization techniques were utilized. In transgenic plants, GUS was expressed from a 2.5-kb endogenous IDD10 promoter. GUS expression was detected in crown roots, lateral root primordia and the vascular region of the seminal root elongation zone (Fig. 2f). To verify GUS expression patterns in root primordia, in situ hybridization was performed with an antisense IDD10 probe. Figure 2(g) shows IDD10 expression in crown roots and primordia of lateral one.

Ammonium uptake and assimilation gene expression in idd10 mutants and IDD10 overexpression lines

Following a period of growth without available N, ammonium will strongly induce many ammonium uptake and N assimilation genes in rice roots. As mutant roots are hypersensitive to ammonium and IDD10 is a nuclear protein with the potential for transcriptional activity, it is possible that IDD10 mediates the expression of ammonium-dependent genes. To investigate this possibility, gene expression was examined in IDD10 mutant and overexpression lines. The control for mutants was a complementation line in which IDD10 cDNA was expressed under its native promoter in an idd10 mutant background. Wild-type siblings segregated from overexpression transgenic lines were the control for overexpressors. After germination, plants were grown in distilled water in a glasshouse for 14 d to ensure the depletion of endosperm nutrients. Seedlings were grown for an additional 3 d in N-free nutrient solution and then transferred to the same nutrient solution containing 0.5 mM (NH4)2SO4. Under these culture conditions, mutant roots did not show any morphological defect. Whole roots were harvested 0, 3, 6 and 12 h after the provision of (NH4)2SO4. Quantitative RT-PCR was performed to measure mRNA levels of ammonium uptake and N assimilation genes in roots. These genes included two low-affinity AMTs, AMT1;1 and AMT1;2, and four assimilation-related genes, cytosolic glutamine synthetase (GS1;2), NADH glutamate synthase 1 (NADH-GOGAT1) and two glutamate dehydrogenases (GDH1 and GDH2), all of which exhibit ammonium induction in roots (Senoda et al., 1990; Tabuchi et al., 2005, 2007). The ammonium-mediated induction kinetics were compared between mutants and controls, and between overexpressors and wild-type siblings (Fig. 3). With the exception of GS1;2 and GDH1, all the genes showed lower induction levels in mutants than in complementation plants (Fig. 3a,c,g,k). In particular, the expression level of AMT1;2 was around one-third lower in mutants than in complementation plants at 3 h after ammonium application (Fig. 3c). Overexpression lines exhibited an approximate induction of two-fold or greater of AMT1;2 and GDH2 than wild-type siblings (Fig. 3d,l). The induction of NADH-GOGAT1 at 0 and 12 h, and the induction of GDH1 at 0 and 3 h, were significantly higher in the OX line than in the wild-type line (Fig. 3h,j). In summary, among the genes examined, IDD10 expression levels had the greatest effect on the induction of AMT1;2 and GDH2, with the magnitude of AMT1;2 induction altering two- to three-fold in both the idd10 mutant and overexpression lines (Fig. 3c,d).

Figure 3.

Ammonium-mediated induction of ammonium transporters (AMTs) and assimilation genes in mutant and overexpression lines. Seventeen-day-old seedlings grown as described in the 'Materials and Methods' section were transferred to nutrient solution containing 0.5 mM (NH4)2SO4. Total roots were sampled at 0, 3, 6 and 12 h after the addition of ammonium. Quantitative real-time polymerase chain reaction (qRT-PCR) of total cellular RNA was used to determine the expression levels of AMTs (AMT1;1, AMT1;2), glutamine synthetase (GS1;2), NADH glutamate synthase 1 (NADH-GOGAT1) and glutamate dehydrogenases (GDH1 and GDH2) in indeterminate domain 10 (idd10) (red bars; a,c,e,g,i,k) and overexpression lines (white bars; b,d,f,h,j,l). Controls are shown by black bars. The control for idd10 was a mutant complementation line in which IDD10 cDNA was expressed under the control of its endogenous promoter. Wild-type siblings from the same line were used for overexpression line controls. idd10 and Com2 lines were generated in Oryza sativa Japonica rice cv Hwayoung, and the OX line and its wild-type siblings were generated in O. sativa Japonica rice cv Dongjin. Com2 indicates complementation line 2. mRNA levels were normalized to that of Ubiquitin mRNA. Error bars represent ± SE of the means from three replicates. Significant differences between samples are shown (*, < 0.05; **, < 0.001; t-test).

IDD10 activates AMT1;2 by binding a specific cis-regulatory element in the promoter

As AMT1;2 is sensitive to IDD10 expression, it is possible that IDD10 acts as a transcriptional activator of AMT1;2. To obtain evidence that IDD10 binds directly to the AMT1;2 promoter, a ChIP assay was performed using a 3-kb promoter region from the AMT1;2 locus and an IDD10-GFP transgenic line. IDD10-GFP lines express IDD10 cDNA fused to a 3′ GFP under control of the Ubiquitin promoter. IDD10:GFP OX lines complemented the idd10 mutation (Fig. S3a). Moreover, the provision of ammonium induced a similar magnitude of AMT1;2 and GDH2 induction to that in IDD10 OX (Fig. S3b,c). Therefore, IDD10-GFP is functionally equivalent to IDD10. Calli derived from IDD10-GFP OX lines were utilized for the ChIP assay. The 3-kb promoter region was divided into six parts, and fragments of c. 120–150 bp from each part were amplified from immunoprecipitates isolated with anti-GFP antibody. One region located between the 151st to 271st bp from the translation start codon was amplified from chromatin immunoprecipitates (Fig. 4a,b). This promoter region was designated as P1. GFP transgenic lines, GFP preimmune serum and a P2 DNA fragment adjacent to P1 were used as negative controls for the ChIP assay. The binding affinity between IDD10 and the P1 fragment was reconfirmed by EMSA using a GST:IDD10 fusion protein (Fig. 4c). Both ChIP and EMSAs indicated the presence of an IDD10 cis-regulatory element in the P1 fragment of the AMT1;2 promoter. To identify the IDD10-binding sequence, the 120-bp P1 fragment was divided into three parts and two additional segments spanning the boundaries of the three parts. EMSAs identified a 40-bp DNA fragment containing an IDD10-binding site (Fig. S4a). From the 40-bp fragment, 30 bp DNA was further separated into three overlapping parts, which finally revealed a 15-bp region responsible for the binding affinity of IDD10 (Fig. S4b). The 15-bp fragment contained 5′-GGACAAA-3′, which is similar to the binding sites of IDD homologs in maize and A. thaliana (ID1, 5′-TTTGTCG/CTTTT-3′ and IDD8, 5′-TTTTGTCC-3′, respectively; Kozaki et al., 2004; Seo et al., 2011a). Inspection of the genomic sequence showed that there is only one GGACAAA motif within the 3-kb promoter region of AMT1;2. To verify that GGACAAA functions as a cis-acting element for IDD10, three DNA fragments were mutated at the putative binding sequence (m1, AGGGAAAAAAATCAA) and its flanking sequences (m2, AAAAGGACAAATCAA; m3, AGGGGGACAAAGGGA). These fragments were examined using EMSA. The data showed that IDD10 was unable to bind to m1, whereas m2 and m3 exhibited slightly higher and lower binding affinities than normal sequences (WT, AGGGGGACAAATCAA), respectively (Fig. 4d). Competition EMSAs were performed between WT and m1, m2 or m3. An increase in unlabeled m1 had little effect on binding between IDD10 and the labeled WT probe, whereas m2 interfered significantly with binding. Therefore, a change in the flanking sequence from GGG to AAA increased the affinity of the core sequences for IDD10 (Fig. S4c). To examine transcriptional activation of the AMT1;2 promoter by IDD10 in vivo, transient expression assays were performed using rice protoplasts (Fig. 4e). Oc cells were co-transformed with pUbi:IDD10 and a vector expressing GUS under either the native (Pwt) or mutated (Pm1) AMT1;2 promoter (2 kb). Pm1 contained the same mutated sequence (AAAAAAA) at the binding site as the m1 probe used for EMSA in Fig. 4d. p35S::Luciferase was used as an internal control. In IDD10-expressing rice Oc cells, GUS activity was around four-fold higher from the native promoter than from the mutated promoter. In summary, these data demonstrate that IDD10 activates the transcription of AMT1;2 by binding a cis-element located 251 bp from the translational start codon.

Figure 4.

Indeterminate domain 10 (IDD10) activates the AMT1;2 promoter by binding a specific cis-regulatory element. (a) The schematic diagram shows the AMT1;2 promoter and the location of probes (P1 and P2) used in the electromobility shift assay (EMSA) and chromatin immunoprecipitation (ChIP) assay. (b) Antiserum against green fluorescent protein (GFP) (Ab) and preimmune serum (Pre) were used for ChIP assays in rice calli expressing IDD10:GFP. P1 and P2 regions were amplified from immunoprecipitated DNA. Input DNA was used as an internal control. (c) EMSAs were performed using the GST:IDD10 fusion protein with P1 and P2. Purified glutathione S-transferase (GST) protein was used as a control. (d) To identify core sequences for IDD10 binding, DNA fragments were mutated at 15 nucleotides (m1, m2 and m3) and utilized for EMSAs. (e) Transient expression assays were performed by the co-transfection of pUbi:IDD10 and an AMT1;2 promoter expressing β-glucuronidase (GUS). Two AMT1;2 promoters were examined: the normal promoter Pwt and the modified promoter Pm1, which contained a mutated IDD10-binding sequence. The IDD10-binding sequence of Pm1 carried AAAAAAA instead of GGACAAA. The 35S promoter-driven luciferase gene was used as an internal control to normalize GUS expression data. Error bars represent ± SE of the means from three replicates. Significant differences between control and effector are shown (*, < 0.001, t-test).

GDH2 is another IDD10 target gene

AMT1;1, GS1;2, NADH-GOGAT1, GDH1 and GDH2 were inspected for the presence of the same sequence motif to which IDD10 binds for the activation of AMT1;2. The GDH2 locus was found to contain two putative IDD10-binding motifs (TTTGTCC/G), called P1 and P2, in the first and fifth introns, respectively (Fig. 5a). As the ammonium response of GDH2 was greatly enhanced in IDD10 overexpression lines (Fig. 3l), the possibility was examined that GDH2 might also be an IDD10 target gene. To determine whether IDD10 binds to these intron sequences, EMSA was performed using 26-bp DNA fragments containing the binding motif sequences. The results showed that IDD10 bound to the P2 fragment, but not to P1 (Fig. 5b). For a P2 control, a mutated fragment (mP2) was used, in which the core sequence was changed from GGACAAA to AAAAAAA. A competition EMSA was performed to verify the binding affinity between IDD10 and the cis-element of the P2 DNA fragment (Fig. 5b). Binding between IDD10 and P2 was reduced significantly in the presence of excess unlabeled P2, but was unaffected by unlabeled mP2. To examine the binding between IDD10 and the P2 region of GDH2 in vivo, a ChIP assay was performed using IDD10-GFP transgenic calli. Genomic DNA was immunoprecipitated with GFP antiserum, and then the P1 and P2 regions were amplified by PCR. Only the P2 region amplified from IDD10-GFP chromatin samples and no P1 DNA was produced from the same immunoprecipitates (Fig. 5c). Pre-immune serum precipitates and GFP transgenic lines were used as controls.

Figure 5.

Indeterminate domain 10 (IDD10) shows the binding affinity for the fifth intron of glutamate dehydrogenase 2 (GDH2). (a) The schematic diagram indicates the locations of two putative IDD10-binding motifs (red ovals) and probes (P1 and P2) used in the electromobility shift assay (EMSA) and chromatin immunoprecipitation (ChIP) assay. (b) P1 and P2 EMSAs of GDH2 introns were performed with a mutated P2 probe (mP2), in which GGACAAA was changed to AAAAAAA in the P2 fragment. Competition EMSAs were performed with 50- and 100-fold excess (50× and 100×) of unlabeled probes. (c) Antiserum against green fluorescent protein (GFP) (Ab) and preimmune serum (Pre) were used for ChIP assays in rice calli expressing IDD10-GFP. The same GDH2 regions used in EMSA were amplified from immune-precipitated DNA. Input DNA was used as an internal control.

IDD10 regulates ammonium-responsive gene expression in roots

To understand how IDD10 expression affects ammonium-mediated gene induction, the microarray data with three biological replicates were analyzed to identify genes with greater than or equal to two-fold (log2 ≥ 1) induction by ammonium. RNAs were sampled from the roots of idd10, IDD10 OX-GFP2 and their wild-type siblings at 0 and 3 h after treatment with 1 mM ammonium, as described above. The wild-type controls for mutants were wild-type siblings from the same line. For overexpressors, wild-type siblings were segregated from the same transgenic lines. Among these ammonium-responsive genes, 348 genes showed higher expression in IDD10-GFP overexpressor lines and somewhat reduced expression in idd10 mutants than in the wild-type siblings. The 1.5-kb promoter sequences of these 348 genes were inspected and putative IDD10-binding sequences (TTTGTC/G) were identified in 175 genes. Table S1 lists the 348 ammonium-responsive genes, together with information on the presence of putative IDD10-binding motifs. Putative IDD10-binding sequences (TTTGTC/G) are present in many ammonium-responsive genes that show IDD10-enhanced expression.

Genes known to play a role in N metabolism or N-linked metabolism were selected from the microarray data and quantitative RT-PCR was used to verify their expression in IDD10 OX2 and IDD10 OX-GFP2. The selected genes encoded the following: two ferredoxin-nitrite reductases (NiRs; Os02g52730 and Os01g25484), plastidic glutamine synthetase 2 (GS2; Os04g56400), trehalose-6-phosphate synthase, cytokinin dehydrogenase and aminotransferase (Os03g48060). Quantitative RT-PCR verified that IDD10 influenced significantly the ammonium induction profiles of these genes (Fig. 6). More than 10 overexpression lines were generated and gene expression was examined in three independent lines with similar results (Fig. S5). The two NiRs, GS2, cytokinin dehydrogenase and trehalose-6-phosphate synthase carry putative IDD10-binding motifs in their promoters or introns. As EMSA showed a strong binding affinity between these sequences and IDD10 in vitro (Fig. S6), it is possible that these motifs function as cis-acting elements in IDD10 transactivation.

Figure 6.

Indeterminate domain 10 (IDD10)-dependent expression of six N-linked metabolic genes. Quantitative real-time polymerase chain reaction (qRT-PCR) was used to verify ammonium-responsive expression detected in the microarray. The selected genes were ferredoxin-nitrite reductase (Os02g53730) (a), ferredoxin-nitrite reductase (Os01g25484) (b), glutamine synthetase 2 (GS2, Os04g56400) (c), aminotransferase (Os03g48060) (d), trehalose-6-phosphate synthase (Os02g54820) (e) and cytokinin dehydrogenase (Os08g35860) (f). Total RNA was extracted from root samples exposed to ammonium for 3 h. Induction ratios were calculated from the ratio between expression in the mutants (black bars) or overexpressor lines (white bars) relative to that in the wild-type controls. Two independent overexpressor lines, OX2 (x) and GFP-OX2 (y), were utilized for qRT-PCR. Ubiquitin was used to normalize the mRNA level of each gene. Data represent mean values ± SE of three replicates.

Effects of MSO and glutamine on IDD10-mediated gene expression

Assimilated ammonium metabolites exert feedback regulation that controls the expression of the OsAMT1 family and N metabolic genes. In rice, glutamine and asparagine can replace ammonium in the activation of these genes and the GS inhibitor MSO prevents the induction of genes involved in N metabolism (Hirose & Yamaya, 1999; Sonoda et al., 2003b). Inorganic N compounds and their metabolites are considered as discrete signaling molecules that trigger N-responsive gene expression (Rawat et al., 1999; Vidal & Gutierrez, 2008). Therefore, it is very important to determine whether IDD10-mediated gene induction is part of the regulatory circuit originating from an inorganic N source or assimilated metabolites. Quantitative RT-PCR was used to determine the effects of MSO and glutamine on IDD10-mediated gene induction. The expression levels of AMT1;2, GDH2, two NiRs and GS2 were measured in total cellular RNAs from idd10, overexpressor and wild-type controls, which were cultured for 3 h in a nutrient solution containing 0.5 mM (NH4)2SO4, a mixture of 1 mM MSO and 0.5 mM (NH4)2SO4, or 5 mM l-glutamine. The concentration of 5 mM of l-glutamine was determined after a series of glutamine levels were examined for their effects on the expression of AMT1;2 and GDH2 (Fig. S7) (Sonoda et al., 2003b). Samples grown in N-free solution were examined as a control. In wild-type plants, the ammonium-mediated induction of AMT1;2 and GDH2 was strongly suppressed by MSO and significantly induced by glutamine (black bars; Fig. 7a–d). By contrast, MSO enhanced the ammonium-mediated induction of both NiR genes, and their expression was reduced by glutamine (black bars; Fig. 7e–h). GS2 expression was reduced slightly by both MSO and glutamine treatments (black bars; Fig 7i,j). These five genes showed similar overall response patterns to MSO and glutamine in idd10 mutants and wild-type plants (red bars; Fig. 7a,c,e,g,i). In the presence of glutamine, mutants exhibited reduced expression of AMT1;2 and GDH2, but enhanced expression of NiR (Os0252730) and GS2. However, it should be noted that, in MSO-treated roots, mutants accumulated more mRNA for GDH2, the two NiRs and GS2 than did the wild-type. In overexpression lines, the five genes showed almost identical response patterns to those of wild-type plants (white bars; Fig. 7b,d,f,h,j). However, these genes produced higher mRNA levels in the overexpression lines than in the wild-type siblings. Overall, it appears that IDD10 was unable to dominate or override the effects of MSO or glutamine on the expression of these genes. Notably, the expression of IDD10 was decreased slightly by exogenous ammonium (Fig. 2e), whereas the expression of IDD10-dependent genes was increased. This implies that there might be a negative feedback mechanism between the expression of IDD10 and the expression of ammonium-responsive genes mediated by IDD10. In conclusion, IDD10 operates in the transcriptional response triggered by N signals and probably functions within regulatory circuits initiated by inorganic or organic N compounds.

Figure 7.

Effects of methionine sulfoximine (MSO) and glutamine on indeterminate domain 10 (IDD10)-mediated gene expression. Seventeen-day-old rice seedlings grown as described in the 'Materials and Methods' section were transferred to medium containing no nitrogen (−N), 0.5 mM (NH4)2SO4, a mixture of 1 mM MSO and 0.5 mM (NH4)2SO4, or 5 mM l-glutamine (Gln). After 3 h, quantitative real-time polymerase chain reaction (qRT-PCR) was used to measure mRNA levels of target genes. Ammonium transporter (AMT1;2), glutamate dehydrogenase 2 (GDH2), two ferredoxin-nitrite reductases (NiRs) and glutamine synthetase 2 (GS2) were amplified from total cellular RNA obtained from the roots of idd10 (red bars; a,c,e,g,i) and IDD10 overexpression (white bars; b,d,f,h,j) plants. Controls for mutants and overexpressors are shown as black bars. mRNA levels were normalized against Ubiquitin mRNA. Data represent mean values ± SE of three replicates. Significant differences between samples are shown (*, < 0.01; **, < 0.05; ***, < 0.001; t-test).


IDD10 as a transcriptional activator

Genes carrying ID zinc finger motifs are present in relatively low copies in plant genomes. Rice and A. thaliana contain c. 15 and 16 genes, respectively (Colasanti et al., 2006). The amino acid sequences of ID domains are highly conserved within the IDD family of proteins, whereas the C-terminal domains are highly diverged. The binding affinity between ID domains and cis-acting DNA sequence elements has been demonstrated in maize ID1 and in A. thaliana IDD8 and IDD14 (Kozaki et al., 2004; Seo et al., 2011a,b). The core sequence of the cis-element identified with IDD10 in this study is similar to the core sequences previously reported from other IDD genes. The IDD family has been demonstrated to be involved in various developmental and metabolic processes during plant development. This study revealed a new function for the IDD gene family. It is likely that the diverse functions of IDD family members might be derived from the specificity of the C-terminal activation domain and from interactions with additional proteins. Indeed, nuclear proteins that bind to IDD proteins have been reported. GRAS family proteins physically interact with MAG, JKD and IDD1 (Welch et al., 2007; Feurtado et al., 2011).

Action mode of IDD10 as a transcriptional activator

In this study, the actions of IDD10 on the expression of target genes have been demonstrated primarily by comparing gene induction magnitudes among knockouts, overexpressors and their wild-type siblings. The induction of the target genes is positively correlated with IDD10 expression, which suggests that the level of IDD10 is a limiting factor for the induction of the target genes. However, the overall induction kinetics of target genes in both mutants and overexpressors are similar to those of their wild-type siblings, although there are substantial differences in expression levels of target genes between mutants and overexpressors. This observation implies that IDD10 has a partial role in the full activation of target genes. A similar effect of IDD10 was observed on feedback regulation mediated by N metabolites. IDD10 affects only the magnitude of the repression or induction of metabolite-responsive target genes. It should be noted that transient transcription assays with rice Oc cell lines show that IDD10 can activate target genes via direct binding to the cis-elements under nonstimulating conditions. Therefore, IDD10 might be necessary not only for N-responsive activation, but also for basal expression of target genes. Considering the fact that these promoters of N-linked genes are controlled by the combined effects of multiple regulators (Konishi & Yanagisawa, 2010; Wang et al., 2010), N-mediated gene induction should be the combined outcome of complex regulatory transcriptional networks. IDD10 is a part of the regulatory circuits. In addition, it is worth noting that IDD10 controls the simultaneous induction of several genes for ammonium uptake and assimilation. This implies that there should be regulatory circuits that co-ordinate the transcriptional induction of N-linked genes.

Ammonium-sensitive root growth

When rice seminal roots are grown in high concentrations of ammonium, they show retarded elongation (Hirano et al., 2008). Ammonium-affected roots frequently develop coiled tips. By contrast, equivalent concentrations of nitrate do not affect significantly root growth. In the presence of the GS inhibitor MSO, exogenous ammonium does not inhibit significantly root growth. Mutant idd10 plants exhibit an almost identical phenotypic response to ammonium, including tip coiling and rescue by MSO. The only difference is that idd10 mutants develop these phenotypes at an ammonium concentration that does not affect root growth of wild-type plants. Thus, idd10 mutants are more sensitive than normal plants to ammonium. These results indicate that the idd10 mutation does not induce a novel ammonium response phenotype, but does induce a hypersensitive phenotype.

Although it is well documented that ammonium induces growth defects in rice plants, the physiological or molecular mechanism that induces the phenotype is unknown. To further characterize the physiological status of seminal roots affected by ammonium, this work examined the internal ammonium levels and the effects of external pH on phenotypic development. Internal ammonium levels were measured in MSO-treated and control roots from mutant and wild-type plants (Fig. S8). In the presence of MSO, mutant roots accumulated more internal ammonium than did wild-type roots. The internal ammonium levels of untreated roots were lower in mutants than in wild-type plants. Therefore, the ammonium hypersensitivity of seminal roots does not correlate with an increase in the internal ammonium levels (Fig. S8). The possibility that mutants might be hypersensitive to acidification of the rhizosphere that is associated with ammonium uptake was examined (Bloom et al., 2003). However, no detectable pH effect on the phenotype was observed in mutants grown in MS salts medium at different pH values (Fig. S9).

Because MSO rescues the ammonium-induced defect of seminal roots, it has been concluded that the assimilation of ammonium is necessary for the development of the root coiling phenotype (Hirano et al., 2008). Our data revealed that, in the mutants, MSO treatment enhanced the expression of some N metabolism genes. It is notable that MSO treatment induced NiRs and GS2 more strongly in idd10 and IDD10 OX than in their wild-type siblings (Fig. 7). A better understanding of the relationship between IDD10 and the regulatory circuits of NiRs and GS2 is required to explain how MSO treatment overrides IDD10-mediated induction of NiRs and GS2. In addition, it remains unknown why the root phenotype is observed in the idd10 mutants in the presence of ammonium. It is possible that this is a result of changes in the expression of IDD10 target genes. Further study is required to identify the genes that are responsible for the hypersensitivity of the idd10 mutants to ammonium.

Possible utility of IDD10 to study N-linked metabolism

Because the cellular levels of metabolites were not analyzed in mutants and overexpressors, the physiological effect of IDD10 on N-linked metabolism and N utilization could not be determined. However, the results suggest that IDD10 could function to control AMT1;2. Within the OsAMT1 family, OsAMT1;1 and OsAMT1;2 are up-regulated following exposure to ammonium, whereas OsAMT1;3 is up-regulated by N deprivation (Kumar et al., 2003; Sonoda et al., 2003a; Suenaga et al., 2003). IDD10 specifically regulates AMT1;2. The transcriptional regulation of the AMT gene might become an important tool in understanding the relationship between ammonium uptake and assimilation genes. As AMT1;2 is extremely sensitive to ammonium in roots, the AMT1;2 promoter and IDD10 could be an excellent combination in developing an ammonium-inducible expression system.

It is important to emphasize that IDD10 influences the expression of the ammonium-generating NiR and ammonium-assimilating GS2, which catalyze two sequential steps in the generation and assimilation of ammonium in plastids. NiR catalyzes the reduction of nitrite to ammonium in plastids. Although NiR is considered to perform a pivotal role in nitrate assimilation, both nitrate and ammonium induce A. thaliana NiR (At2g15620) in roots (Patterson et al., 2010). The rice genes with the greatest homology to the A. thaliana NiR gene are Os02g52730 and Os01g25484, which were identified as IDD10 target genes in this study. The amino acid sequence of AtNiR exhibits 71% and 73% identity to the OsNiR proteins encoded by Os02g52730 and Os01g25484, respectively. NRE (nitrate-responsive cis-element) is a cis-acting sequence for nitrate induction that has been identified in an A. thaliana NiR promoter. NRE is sensitive to nitrate, but insensitive to exogenous glutamine. Inspection of the NRE sequence revealed that there was no putative cis-element with binding affinity for IDD10 or any IDD proteins previously reported in A. thaliana or maize. To compare the frequency of the IDD10 cis-element with that of a nitrate-responsive cis-element (tGaCCTTT-N(9-10) – AAGaG), the 1.5-kb promoter regions of the top 50 ammonium-induced genes identified in the microarray analysis were examined. The promoter regions of 18 genes contained IDD10 cis-elements, whereas no promoter regions contained a nitrate-responsive cis-element. Plastidic GS2 is encoded by a single gene and is primarily responsible for the assimilation of ammonia arising from photorespiration and nitrate reduction in photosynthetic tissues. GS2 is also expressed in roots (Zozaya-Hinchliffe et al., 2005; Zhao & Shi, 2006). In rice, as in other plants, the expression of GS2 is more sensitive to nitrate than to ammonium. Therefore, it is important that IDD10-mediated regulation of N-linked genes is also examined in shoots.

Trehalose biosynthesis, the pentose phosphate pathway and the cytokinin response are over-represented functional categories among inorganic N-induced gene sets (Mok & Mok, 2001; Wang et al., 2003; Scheible et al., 2004). The fact that IDD10 enhances the expression of genes for aminotransferase, trehalose-6-phosphate synthase and cytokinin dehydrogenase further strengthens the possibility of a functional involvement of the gene in N-linked metabolism. Therefore, it will be important to perform a metabolite analysis to estimate the possible agricultural benefits of manipulating IDD10 to improve the efficiency of N metabolism in crop plants.


This work was supported by grants from the Next-Generation BioGreen 21 Program (PJ008215 and PJ008168), Rural Development Administration, South Korea. We are grateful to Dr. Gynheung An (KyungHee University, South Korea) for the T-DNA insertion mutant. Y.H.X, J.M.L and R.A.P. were supported by a scholarship from the BK21 program.